Air traffic management and weather: the potential of an ...T2B... · DIESER TEXT DIENT DER...

DIESER TEXT DIENT DER NAVIGATION

Air traffic management and weather: the potential of an integrated approach

WMO WWRP 4th International Symposium on Nowcasting and Very-short-range Forecast 201625-29 July 2016, Hong Kong

Martin Steinheimer, Carlos Gonzaga-Lopez, Christian Kern, Markus Kerschbaum, Lukas Strauss

Kurt Eschbacher, Martin Mayr, Carl-Herbert Rokitansky

Funded by:

Outline

The MotivationWhy are we doing it…

The ProjectWho is doing it…

The MethodHow are we doing it…

The ToolWhat do we use to do it…

The First ResultsWhat we did so far…

The OutlookWhat are we doing next…

2

3

The Motivation

Weather impact on Air Traffic Management

Weather especially wind, thunderstorms and low visibility have big impact on airport capacity

Weather cannot be changed but accurate forecasts help to be prepared and to minimize weather impact

Weather impact in numbers:– Vienna International airport:

– Winter 2015/16 weather delays mainly due to low visibility (almost no snow at Vienna airport)

4

Weather impact on Air Traffic ManagementLow Visibility Procedures

What are Low Visibility Procedures

LVP seen from the cockpit: https://www.youtube.com/watch?v=mSNE3SmYA-8

5

LVP state RVR Ceiling Separation Capacitynormal 2.5NM >40

LVP < 600m BKN < 200ft 4NM 25LVP CATIII < 350m 6NM 18

or

6

The Project

MET4LOWW – research projectMET potentials in arrival and departure management

Funded by the Austrian Research Promotion Agency (FFG)

Participants– Austro Control (ATM and MET department)– Uni Salzburg, Aerospace Research Group– DLR Institute of Atmospheric Physics

Objective: Evaluate the potential of a holistic ATM/MET approach:– Final approach

Time Based Separation (pair-wise/weather dependant separation)Low Visibility ProceduresWind shifts (=RWY direction changes)

– Arrival managementThunderstorms

– Departure managementMET input to Airport Collaborative Decision Making

7

8

The Method

Weather impact analysisFlow chart

9

NAVSIM KPIs Impact Analysis

ATM Measures

(LVP, WV, dual RWY,…)

Wea

ther

OBS

FCST

ATM procedure

(LVP, WV, dual RWY,…)

Traf

fic

Flight-Plan(demand)

• Low Visibility Procedures• Wind• Thunderstorms

• Traffic regulations• ATCO staffing• short-term measures

• Key Performance Areas• Key Performance Indicators• Derived Economic Value

• Separation on final approach• RWY in use• Traffic routing

• Air Traffic Simulation

• Real traffic• Generated traffic

Weather impact analysisKPAs / KPIs

Key Performance Areas considered– Capacity– Environmental Impact / Flight efficiency– Cost-effectiveness– Traffic complexity

As proxy for Safety (safety can not be measured with perfect simulations)

Each KPA is represented by one or more Key Performance Indicators, which should meet following criteria:– Specific– Measurable– Drive the desired behaviour– Accountable/manageable– Compatible with ICAO guidelines– Proper with regard to weather forecasts

10

According to EUROCONTROL 2011 technical note: Measuring Operational ANS performance at Airports

Weather impact analysisForecast value

Following a similar approach to using contingency table and cost matrix

A contingency table and a KPI matrix can be used to assess the forecast value

11

Observed

Yes No

ForecastedYes hit false

alarm

No missed Correct negative

o = h + m 1 - o

Observed

Yes No

Take action

Yes C + L – L1 C

No L 0

Observed

Yes No

Take action

Yes KPIh KPIf

No KPIm KPIn

Observed

Yes No

ForecastedYes hit false

alarm

No missed Correct negative

o = h + m 1 - o

(e.g.: Richardson, D. S., 2000: Skill and relative economic value of the ECMWF ensemble prediction system. Q.J.R. Meteorol. Soc., 126, pp. 649-667.)

Weather impact analysisForecast value

The KPI matrix can be filled using the air traffic simulator

12

Observed

Yes No

Take action

Yes KPIh KPIf

No KPIm KPIn

Wea

ther

eve

nt

occu

rs in

si

mul

atio

n

Wea

ther

eve

nt

does

NO

T oc

cur i

n si

mul

atio

nReduced traffic

according WX-FCST

full traffic

Weather impact analysisForecast value

Cost/Loss ratio can be derived from the KPI matrix Economic value can be derived from contingency table and KPI-matrix

(similar to potential economic value)

13

NOTE: probabilistic curve is not based on actual verification results C/L derived from estimates not actual evaluation

(2h LVP forecasts at LOWW in 2015)

NOTE: KPI values and C/L derived from estimates not actual evaluation

Weather impact analysisChallenges

KPIs contradictory, e.g.:– trade-off between maximizing capacity and optimizing workload– trade-off between optimizing workload and minimizing flight delays– etc…

Different stakeholders (ANSP, airlines, airports,…) associate different priorities to KPAs/KPIs

– e.g. ATM workload is not airlines’ first priority In order to quantify the impact on the overall air traffic

management system:– The various KPAs/KPIs need to be combined– That requires appropriate normalization and weighting considering

all stakeholders’ requirements A detailed analysis on this topic was done in an Eurocontrol

commissioned research study(Bert De Reyck, B., Degraeve, Z. and Grushka-Cokayne, Y., 2006: Decision Support Using

Performance Driven Trade-Off Analysis. EEC Note: EEC/SEE/2006/001)

14

15

The Tool

Building Blocks – NAVSIMATM/ATC/CNS Simulator – Main Characteristics

Detailed world-wide Runway-to-Runway (or Gate-to-Gate) Air Traffic Simulation

Using detailed Aircraft Performances Based on around 1 million Nav-data (as used for FMS) Using sophisticated Simulation Techniques Simulate more than 10.000 Aircrafts (AC) simultaneously – generic FMS

for each AC Displaying today's and any future predicted Air Traffic Simulation running in Real time or Fast Time mode Inclusion of third party test equipment and products Supports Evaluation of NextGen / SESAR concepts

16

Note: NAVSIM ATM/ATC/CNS Tool developed by Mobile Communications Research & Development ForschungsGmbH in co-operation with USBG

Building Blocks – NAVSIMArchitecture

17

INPUT Data

Navigation

Flightplans

Performances

Meteorological

Configuration

Script of Scenario

Live Data

SimulationProcess

Radar Display

Aircraft Mobility

FMS Function

ATS/AOC Comms

Display Control

Speed Control

Area Select

Interactive

Mode

OUTPUT Data

Visualisation

Generation

Evaluation

Verification

Validation

Frequency Planning

Interfaces

Other

NAVSIM / AMAN Advanced Arrival Management

NAVSIM/AMAN Advanced Arrival Manager includes the following features & functions for ATC performance analysis and evaluation of MET-potentials: Detailed Arrival Management of all aircraft (starting calculation about 200 nM to

80 nM ahead of destination aerodrome at "entry point") Detailed Merge Point Calculation (e.g. IF or Final Approach Fix FAF) overfly

time based on 3 basic modes: Direct Mode (no transition required), Transition Mode and Holding Mode (if required)

For each flight the flight path geometry, length and Calculated Time of Arrival(CTA) is computed at entry point and remains stable (unless adjustments to flight behavior and or current weather situation becomes necessary) until touch down on arrival runway

Continuous Descend Approach (CDA, glide slope 3 degrees) is calculated and executed at entry point

For Wake Turbulence calculation for each aircraft type the wake categoryaccording to ICAO rules or (new) RECAT rules is assigned and taken into account during Departures, within TMA and on Final Approach

18

NAVSIM / AMAN Advanced Arrival Management

Distance Based or Time Based Minimum Separation on Final Approach are selectable and taken into account in Arrival Management calculations

Low visibility procedures (increased Minimum Separation distances or time) are taken into account in all Arrival Management calculations

Wind profiles per runway / within area taken into account Optimized Weather Avoidance path is calculated (based on Current and

Nowcast data) in case of adverse weather (CBs) Harmonization between departing and arriving air traffic is taken into account

by NAVSIM/AMAN Synchronous arrivals on parallel and/or crossing runways are possible NAVSIM/AMAN allows comparison between optimized flights and “best

practices” based ATCO controlled flights (based on track/CPR data) Detailed performance analysis (in terms of Key Performace Indicators (KPIs))

are calculated & recorded of optimized NAVSIM/AMAN calculated flight paths / time taking all of the above rules and features into account !

19

NAVSIM / AMANValidation

20

Compare actual flight path to simulated flight path– Simulation is initialized with actual traffic at STAR endpoints– Compare simulation and actual flight paths between STAR endpoints and touchdown

direct modetransition modeholding mode

NAVSIM / AMANValidation

21yellow: CPR; blue: simulated

Validation experiments show very good agreement between simulation and actual flight tracks– ATCOs certify widely realistic behaviour of simulator

Vienna Airport

PESAT

BALAD

MABOD

NERDUholdings

KPI statistics

simulated

actual

22

NAVSIM / AMANValidation - video

Low Visibility Procedures (LVP) during morning rush hour

NAVSIM / AMANValidation - video

Low Visibility Procedures (LVP) during morning rush hour

23yellow: CPR; blue: simulated

Validation result:It is reasonable to use the simulator for weather impact evaluation experiments

24

The First Results

Case studiesLow Visibility Procedures

Simulation of two scenarios– Short period (1.5h) of LVP

during morning peak

25

– Long period (13h) of LVP during daytime

Case studiesLow Visibility Procedures

For both scenarios multiple variations (weather + traffic) were simulated:– n: No LVP observed and none forecast (i.e. full traffic)– f: No LVP observed, but forecasted (i.e. traffic regulated)– m: LVP observed, but not forecasted. Traffic is regulated

once LVP observed– h: LVP observed and forecasted. Traffic regulated according

forecast.

26

Observed

Yes No

Take action

Yes h f

No m n

Case studiesLow Visibility Procedures - KPIs

n f m hTrackmiles / flight 62.4 60.8 69.2 65.4

Holding time [min] 52 18 327 94

Holding time / flight [min] 0.17 0.06 1.04 0.30

Delay [min] 0 899 3744 5594

Delay / flight [min] 0 2.9 11.9 17.9

ATCO phrases 3159 3076 3515 3236

n f m hTrackmiles / flight 60.1 59.5 59.8 58.8

Holding time [min] 8 10 6 4

Holding time / flight [min] 0.08 0.10 0.06 0.04

Delay [min] 0 175 215 187

Delay / flight [min] 0 1.7 2.1 1.8

ATCO phrases 1069 1074 1069 1065

Short event:1.5 hours103 flights

Long event:13 hours314 flights

Case studiesLow Visibility Procedures

Temporal variation of KPIs must be considered too– e.g. Air Traffic Control Officer commands as a measure of frequency

occupation

28

29

The Outlook

OutlookWhat are we doing next…

Refine impact analysis Run simulations of other weather dependant scenarios

– Time Based Separation / pairwise separation / weather dependant separation (incl. DLR Wake Vortex-model)

– thunderstorms in approach sectors Derive potential for optimized holistic ATM/MET procedures

– How can weather forecasts be improved– What is the potential of using probability forecasts– Can ATM procedures be adapted to make better use of the

forecasts Extensive validation and evaluation incorporating Air Traffic

Controllers Validation workshops with other stakeholders Possible collaboration with other Air Navigation Service

Provider and aviation MET services

30

Air Traffic SimulatorProf. Dr. Carl-Herbert RokitanskyComputer Sciences Department / Aerospace ResearchUniversity of SalzburgEmail: roki@cosy.sbg.ac.atWeb-Info: www.aero.sbg.ac.at

MET + ATM EvaluationDr. Martin SteinheimerMET Development and InnovationAustro Control GmbHEmail: martin.steinheimer@austrocontrol.at

31

Contact:

Any questionsor comments

32

Additional Information

NAVSIMMET4LOWW – AMAN/DMAN Optimization

Research topics:– Human-in-the-loop simulation of MET4LOWW TMA operations.– Evaluation of MET4LOWW optimization support tools.

33

Image Frequentis.

NAVSIMData Exchange

All tools interconnected by SWIM-like XML protocol (X23)– USBG's distributed human-in-the-loop simulation environment at

the "Aviation Competence Center Salzburg" (ACCS) can also include 3rd party functions and tools (via TCP/IP)

34

Air Traffic Generator

(Navsim TG)

Controller HMI (Navsim ATC)

Flight & AirportObjects

Flight & AirportObjects

Human Controller(s)

Simulated Air Traffic

Controller HMI (SmartStrips)

FlightObjects

Commands

Commands

Controller HMI (Auxiliary Screen)

Flight & AirportObjects

DA42

Human Pilot(s)

Tower visual-ization (XPL10 )

Flight & AirportObjects

Virtual Tower

X23

SWIM

Translation

SWIMNode